In this proof of concept, I share how I built a local RAG system using Qwen1.5 (0.5B), Ollama and LangChain, a step-by-step pipeline to query your documents and understand how it works.
This is great for start. Also, Do we have similar apis for Java ?
Thanks Pritam! Yes, but not end-to-end in Java yet
The best approach now is to use Python + LangChain to handle embeddings, chunking, and vector DB setup (Qdrant, Weaviate, etc.).
Then, your Java app just queries the vector DB and calls the LLM (via Ollama’s HTTP API).
Cleaner separation, better tools, and no need to force everything into Java.
Thanks for sharing . Great way to get hands-on with RAG
Thanks a lot, Jawhar!
This is great for start. Also, Do we have similar apis for Java ?
Thanks Pritam! Yes, but not end-to-end in Java yet
The best approach now is to use Python + LangChain to handle embeddings, chunking, and vector DB setup (Qdrant, Weaviate, etc.).
Then, your Java app just queries the vector DB and calls the LLM (via Ollama’s HTTP API).
Cleaner separation, better tools, and no need to force everything into Java.
Thanks for sharing . Great way to get hands-on with RAG
Thanks a lot, Jawhar!